Adjusting for selection bias in assessing the relationship between sibship size and cognitive performance
2015 (English)In: Journal of the Royal Statistical Society: Series A (Statistics in Society), ISSN 0964-1998, E-ISSN 1467-985X, Vol. 178, no 4, 22 p.925-944 p.Article in journal (Refereed) Published
Consistent negative correlations between sibship size and cognitive performance (as measured by IQ and other mental aptitude tests) have been observed in past empirical studies. However, the issue of potential selection process in the decision to have larger families (largersibship size) has been partly neglected in past single- and multilevel investigations. The present work extends existing knowledge in three aspects: (1) as factors affecting decision to increase family size may vary across the number and composition of current family size, we propose a sequential probit model (as opposed to binary or ordered models) for the propensity to increase sibship size; (2) we investigate if families who choose to increase family size are a representative random sample of the population of families or there exists selection; (3) in order to disentangle selection and causality we propose a multilevel multiprocess modelling where a continuous model for performance is estimated jointly with a sequential probit model for family-size decisions. The issues are illustrated through analyses of scores on PIAT tests among children of the NLSY79. We found substiantial between-family heterogeneity in the propensity to increase family size - thereby providing empirical evidence in support of the admixture hypothesis. Ignoring such adverse selection led to overestimation of the negative effects of sibship size on cognitive performance but our multiprocess modelling could mitigate the biasing effects of selection.
Place, publisher, year, edition, pages
2015. Vol. 178, no 4, 22 p.925-944 p.
Cognitive performance, Multilevel multiprocess modelling, National Longitudinal Survey of Youth 1979, Selection bias, Sequential modelling, Sibship size
Probability Theory and Statistics
IdentifiersURN: urn:nbn:se:su:diva-46174DOI: 10.1111/rssa.12098ISI: 000365391100008OAI: oai:DiVA.org:su-46174DiVA: diva2:371901